Rapid Responses to:

EDITORIALS: Karin J Denton Liquid based cytology in cervical cancer screening BMJ 2007; 335: 1-2 [Full text]

Rapid Responses: Submit a response to this article

Rapid Responses published:

It's the samping, stupid

Paul A O'Brien   (21 July 2007)

It's the samping, stupid 21 July 2007
Paul A O'Brien, Associated Specialist Westside Contraceptive Services, Westminster PCT Send response to journal: Re: It's the samping, stupid	In her recent editorial on liquid-based (LBC) and conventional cytology (1), Denton fails to address what is probably the key remaining challenge for the cervical screening programme, the poor sensitivity of both of these technologies. In the furore that followed the Leicestershire cervical cancer audit (2) the most important message was lost. Of the 324 women who had a smear prior to diagnosis, 62% (202) had a correctly assessed normal smear. There was no sign of pre-cancer on the slide. Laboratory error accounted for only one third of the false negatives tests. If LBC, in itself, will do nothing to solve this problem. The abnormal cells have to get into the pot. Cytology labs in this country were already operating to high standards, partly through the routine audit of slides. Improvement in screening will come though sampling technique rather than cytology. As Bill Clinton would say, ‘It’s the sampling, stupid’. But there is no one auditing my practice. If I sample just half of the transformation zone, neither the lab nor anybody else will ever know. The confusion of the source of error in cervical screening can be seen in the leaflet for patients from the NHS Cervical Screening Programme (NHSCSP), Cervical Screening – the Facts. Patients are told that the reliability of the test can be caused by abnormal cells on the slide not being recognised, but nothing on the main contribution to the poor detection rate, the problem of collection. The recent publication Audit of Invasive Cervical Cancers from NHSCSP, which defines a national protocol for audit of cases of invasive cervical cancer, shows how little attention is paid to sampling in audit. The NHSCSP introduced a training programme for smear takers to, amongst other things, reduce the sampling error. I fear that the progress being made in sampling the cervix may be lost as the focus shifts to LBC and HPV typing and vaccination. When a woman walks into the clinic for a smear the single most important thing for her is that if she has a lesion on the cervix, we will detect it. But the sensitivity of a single test is low, probably in the region of 50% for all pre-cancerous lesions and around 70% for high grade lesions (3). Unfortunately the Health Technology Assessment and review by the National Institute for Clinical Excellence of LBC (4) was based mostly on discordant cytology, and mostly in high-risk patients. No attempt was made to assess the true sensitivity of either technology in routine practice, which requires the gold standard of colposcopy/histology. This reluctance to consider the true error rate continues. Despite the attempts of the NHSCSP training programme for smear taker to improve the detection of cervical pre-cancer, many lesions are undoubtedly missed because of inadequate knowledge, a sub-optimal sampler and poor technique. The next time you are going for a smear ask the nurse, or doctor, what is the transformation zone? This is the target for cervical sampling. Many will give the wrong answer. Then ask them how they will sample it? They will tell you that they will rotate the cervical brush into the cervical canal and rotate 5 times in a clockwise direction, in the assumption that that will give an adequate sample. But watch what happens to the bristles of the brush in the British Society for Clinical Cytology training video and the ThinPrep training video – the bristles converge towards the cervical os, and not splay out as is intended by the design, routinely missing much of the transformation zone. Furthermore, the instruction to rotate in a clockwise direction only has value for sampling the cervical canal. Once the bristles bend on the ectocervix, the sharp edge is lost and the direction is irrelevant as the bristles glide over, rather than scrape, the surface of the ectocervix. And ask her how she samples in the presence of an ectropion. Then decide if your pre- cancer is likely to be detected. The Leicestershire audit suggests there is a huge potential for the cervical screening programme to reduce the incidence of cervical cancer by improvements in sampling. What can be done to improve sampling? * Publish the audits of cervical cancer that are being done throughout the country, and determine the proportion of cancers due to inadequate sampling * Intensify the training programme for smear takers * Commission, and publish on the web, videos showing the correct techniques for sampling for the whole range of types and location of transformation zone, and require all smear takers to watch * Routinely provide plastic spatulas to all testers to sample the ectocervix * Perform a randomised trial of quality training in cervical sampling * Promote the development of improved collection devices and test in a randomised trial * Re-audit cervical cancers for the proportion missed because of inadequate sampling It will take a number of years before improved training and skill will impact on the overall number of abnormal tests. The increased sensitivity of the test will increase the cost of screening caused by the relative high prevalence of low-grade lesions. By this time, HPV typing may detect the lesions that are likely to progress. It would seem that the main issue to be addressed in cervical screening is the challenge of sampling. Until we have a device that is effective in sampling the transformation zone and is simple to use, the main improvement in detection of pre-cancer will come from a determination to improve the training of smear takers. Maybe it is time for a clinician who takes routine smears to lead the quality assurance programme. References: 1. Denton KJ. Liquid based cytology in cervical cancer screening. bmj 2007;335(7609):1-2. 2. Symonds P, Naftalin N, Shaw P. A smear on audit. Implications of the Leicester cervical smear audit. bjog 2003;110(7):646-8. 3. McCrory DC, Matchar DB, Bastian L, Datta S, Hasselblad V, Hickey J, et al. Evaluation of cervical cytology. Evid Rep Technol Assess (Summ) 1999(5):1-6. 4. National Institute for Clinical Excellence. Guidance on the use of liquid-based cytology for cervical screening. Technology Appraisal 69. London. 2003. Competing interests: None declared